Re: [Pkg-octave-devel] [RFU] octave-octgpr 1.2.0-3
* Thomas Weber <tweber@debian.org> [2012-03-24 17:01]:
> On Thu, Mar 22, 2012 at 10:12:37PM +0100, Rafael Laboissiere wrote:
> > Please, upload it.
>
> Does running the demo as test makes sense?
Maybe not. I only included it in the Debian package because there are
no tests in the octgpr package.
> ==========================================================
> [demo_octgpr]
> a peaked surface
> warning: print.m: ghostscript not found in PATH
> warning: print.m: Ghostscript binary is not available.
> Only eps output is available.
> sampled at 150 random points
> GPR model with heuristic hypers
> GPR model with MLE training
> ^Mevals: 0 -log(lhood): 1.79769e+308 ^Mevals: 1 -log(lhood): 2.72843e+02 ^Mevals: 2 -log(lhood): 9.82815e+00 ^Mevals: 3 -log(lhood): -5.47395e+01 ^Mevals: 4 -log(lhood): -8.98639e+01 ^Mevals: 5 -log(lhood): -8.98639e+01 ^Mevals: 6 -log(lhood): -1.50112e+02 ^Mevals: 7 -log(lhood): -1.50112e+02 ^Mevals: 8 -log(lhood): -1.51416e+02 ^Mevals: 9 -log(lhood): -1.51722e+02
> converged.
> slightly clustered random points k-means iteration 1^Mk-means iteration 2^Mk-means iteration 3^Mk-means iteration 4^Mk-means iteration 5^Mk-means iteration 6^Mk-means iteration 7^M
> ==========================================================
>
> I'm not sure what to make of the output; is this good or bad?
In the case above, it means that the test succeeded. Keeping this "test"
in the Debian package would make us able to detect eventual run-time
problems in the future.
It is your call to keep or to drop debian/check.m. Either way is okay
with me.
Rafael
Reply to: